See accompanying article by Fox et al. DOI: 10.1002/biot.201100209
Version of Record online: 2 MAR 2012
Copyright © 2012 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
Special Issue: Frontiers in Biochemical Engineering
Volume 7, Issue 3, pages 328–329, March 2012
Jin, Y.-S. and Cate, J. H. D. (2012), Model-guided strain improvement: Simultaneous hydrolysis and co-fermentation of cellulosic sugars. Biotechnology Journal, 7: 328–329. doi: 10.1002/biot.201100489
See accompanying article by Fox et al. DOI: 10.1002/biot.201100209
The production of ethanol from starch or sugarcane-based substrates has been practiced at large scales for decades. Yet, despite tremendous effort, no economical production of ethanol from the abundant plant biomass in plant cell walls has been implemented. Two of the many outstanding problems in the cellulosic ethanol industry that must be resolved include the economic saccharification of recalcitrant plant cell walls and the efficient fermentation of mixed sugars (glucose and xylose) prevalent in cellulosic hydrolysates. In other words, the industry needs better celluloytic enzymes and better fermenting microbes to produce cellulosic ethanol. In this issue of Biotechnology Journal, a comprehensive model of cellulose hydrolysis and fermentation of prevalent sugars in lignocellulosic hydrolysates has been developed to evaluate the performance of cellulase enzymes and fermenting microbes in various ethanol production scenarios. This model will prove useful in guiding the engineering of cellulase enzymes and fermenting microbes for producing cellulosic ethanol economically.
As endo and exo-type cellulase enzymes are strongly inhibited by intermediate (cellodextin) and final products (glucose) of the hydrolysis reaction, efficient and timely removal of these inhibitory products via fermentation in simultaneous saccharification and fermentation (SSF) is a promising process for producing cellulosic ethanol. Historically, the crucial enzyme in cellulase cocktails that generate glucose from cellodextrins, β-glucosidase, must be supplemented due to relatively weak β-glucosidase activities of most fungal cellulase mixtures, and the fact that Saccharomyces cerevisiae, the workhorse yeast for ethanol production, cannot use cellodextrins such as cellobiose as a carbon source, but is superb at fermenting glucose. The recent discovery of cellodextrin transporters and intracellular β-glucosidase from the cellulolytic fungus Neurospora crassa made it possible to engineer S. cerevisiae to ferment cellobiose directly . In addition, this direct utilization of cellobiose by engineered S. cerevisiae opened the possibility of simultaneous co-fermentation of cellobiose and xylose [2, 3]. By improving ethanol yields and productivities from the synergy between cellobiose and xylose fermentation and reducing usage and cost of cellulase enzymes, the co-fermentation of cellobiose and xylose could be an alternative route for producing cellulosic ethanol economically.
Fox et al. , in this issue of Biotechnology Journal, report a comprehensive mechanistic model capturing (i) molecular interactions between individual cellulases and glucan chain lengths, which are dynamically changing during the hydrolysis of cellulose rather than assuming simplified equilibriums for cellulase-cellulose complexes, (ii) known inhibitory effects of glucose, xylose, and ethanol on cellulase enzymes, (iii) differential activities of cellulases based on enzyme location on cellulose and on glucan chain lengths, and (iv) yeast growth and fermentation rates on glucose, cellobiose, and xylose, incorporating glucose inhibition of cellobiose and xylose utilization and ethanol inhibition of cell growth (Fig. 1A). This is not only the most comprehensive mechanistic model of cellulosic ethanol production, but also the first model describing co-fermentation of cellobiose and xylose by engineered S. cerevisiae.
To evaluate the performance of a prototype yeast strain (DA24-16) that encodes a cellodextrin transporter, an intracellular β-glucosidase, and a xylose utilization pathway as well as being capable of co-fermenting cellobiose and xylose simultaneously , Fox et al.  simulated various SSF and separate hydrolysis and fermentation (SHF) conditions with a large number of empirical parameters. Using an extremely high Monod constant (Kcellobiose = 350 mM, ∼120 g cellobiose/L) for growth on cellobiose and a low starting inoculum of yeast cells (OD600 ∼1.0), the model predicted that the co-fermenting strain of yeast produces ethanol more slowly from a mixture of cellulose and xylose than a control case where β-glucosidase was added extracellularly to a strain exhibiting the same glucose and xylose fermenting capability (Fig. 1B). This prediction makes sense, as a much lower Michaelis constant (KM = 0.88 mM, ∼0.3 g cellobiose/L) for β-glucosidase was used for the comparison; however, the co-fermenting strain was able to produce ethanol almost as fast as the control case under SSF conditions when a reduced Monod constant (Kcellobiose = 1.7 mM, ∼0.6 g cellobiose/L) for growth on cellobiose is used (Fig. 1B). These predictions clearly show the importance of efficient elimination of cellobiose for SSF and suggest an effective strategy to improve the co-fermenting strain. In order for co-fermenting strains that utilize intracellular β-glucosidase to be competitive, it will be necessary to reduce the Monod constant for growth on cellobiose.
While the mechanistic model-based predictions of Fox et al.  provide a quantitative means for understanding the dynamics of cellulose hydrolysis and fermentation in SSF scenarios and for improving yeast strains for commercial processes, there are always difficulties in extrapolating from the near ideal conditions used to derive the empirical parameters to the unknowns in a heterogeneous (solid/liquid) and multi-component (cellulose, cellobiose, enzymes, yeast cells, and ethanol) system. For example, cellulases likely do not behave as modeled in high cellulose loadings (>10% weight/volume) due to unknown mechanisms that result in lower hydrolysis yields . A second example is the likely inoculum of yeast used in fermentations, which may be quite high in industrial fermentation . Additionally, the effect of ethanol inhibition on β-glucosidase activity needs to be evaluated. The model developed by Fox et al.  will now allow for a “virtuous cycle” of comparing experimental results to a quantitative model and updating the parameters used. It will demonstrate its greatest benefit as the parameter values are further refined to reflect new empirical measurements of the performance of engineered strains and proposed processes for cellulosic ethanol.
Simultaneous co-fermentation of cellodextrins and xylose could potentially lower the cost of lignocellulosic biofuel production enough to be competitive with fossil fuels. But there are many other steps in the process that need to be addressed to finally achieve this goal. Advances in feedstock harvesting, storage, pretreatment, and fuel separation all will be needed for cellulosic biofuels to finally become cost competitive.
This work was supported by the Energy Biosciences Institute.
The authors declare no conflict of interest.